Learning for Living

Pemodelan dan Analisis Harga Saham – General Electric, Co.

2014-08-16T12:00:00.001+07:00

Halo.

Tulisan ini mendiskusikan dasar teori pemodelan harga saham dan imbal hasil untuk perusahaan General Electric. Dari dasar teori, akan dilakukan pemodelan imbal hasil untuk periode 2006 – 2011, dengan proses rerata dan proses varian, model simetris dan asimetris. Dapat ditunjukkan bahwa periode 2008 – 2009 merupakan periode dengan volatilitas sangat tinggi dibandingkan dengan periode lainnya –dengan waktu-paruh persistensi selama 51 hari. Selain itu, pada periode yang dianalisis menunjukkan adanya leverage yang signifikan dan informasi negatif yang muncul di pasar akan menyebabkan gejolak harga saham yang lebih besar dibandingkan masuknya informasi positif. Lalu, dengan menggunakan model asimetris, validitasnya akan diuji dengan cara membandingkan antara pergerakan saham ramalan dan aktual.

Stasioneritas Data dan White Noise

Syarat pertama suatu deret waktu dapat dimodelkan dalam teori ekonometri adalah deret tersebut harus memiliki sifat stasioner. Deret stasioner ditandai dengan ciri-ciri; rerata deret konstan E(y_t) =μ, varian deret konstan E(yt − μ)(yt − μ) = σ² < ∞, dan autokovarian yang juga konstan E(y_t1 − μ)(y_t2 − μ) = γ_t2−t1.

Uji Unit Root, Augmented Dickey-Fuller pada persamaan

y_t = φy_t−1 + u_tPersamaan (1)

dengan H₀: φ = 1 dan H₁: φ < 1, dapat digunakan untuk menentukan apakah suatu deret waktu stasioner atau tidak. Bila nilai t-statistic-nya lebih kecil daripada t-critical (nilai ini tergantung pemilihan tingkat keyakinan α), maka hipotesa null dapat ditolak; data sudah stasioner. Selanjutnya, akan dilakukan diferensiasi pada deret tersebut sekali atau lebih hingga diperoleh deret yang stasioner, untuk kemudian dapat masuk pada tahap pemodelan.

Konsep lain yang cukup penting adalah deret yang white noise. Ciri-cirinya adalah rerata dan varian yang konstan E(y_t) = μ, var(y_t ) = σ², dan deret tersebut tidak autokovarian, kecuali pada lag nol γ_t−r = σ² bila t = r, selain itu γ_t−r = 0. Agar suatu pemodelan dapat digunakan dalam analisis dan pengambilan kesimpulan, residu dari model tersebut harus memiliki sifat white noise.

Model Regresi Klasik

Pada prinsipnya, struktur persamaan yang diberikan oleh [1] dan digunakan dalam analisis kali ini adalah

y_t = β₁ + β₃x_3t + · · · +β_k x_kt + u_t dimana u_t ~ N(0,σ²) Persamaan (2)

atau dalam bentuk matrik

y = Xβ + u Persamaan (3)

dimana y berukuran T x 1, X berukuran T x k, β berukuran k x 1, dan u berukuran T x 1. Selanjutnya, koefisien estimasi β dihitung dengan persamaan

Persamaan (4)

Untuk kemudahan, perangkat lunak EViews akan digunakan untuk melakukan estimasi parameter seperti pada persamaan (4) diatas.

**Proses Autoregresive (AR) dan Moving Average (MA) dan Metodologi Box-Jenkins**

Kelas pemodelan deret waktu yang paling sederhana adalah proses moving average. Misalkan u₁ (t = 1,2,3,…) merupakan proses white noise dengan E(ut) = 0 dan var(ut) = σ². Maka,

y_t = μ + u_t + θ₁u_t−1 + θ₂u_t−2 +· · ·+θ_qu_t−q Persamaan (5)

merupakan persamaan moving average orde ke-q dan disimbolkan dengan MA(q) [1]. Persamaan ini menggambarkan bahwa nilai suatu deret pada waktu tertentu memiliki rerata sebesar μ dan merupakan fungsi galat white noise pada lags sebelumnya.

Disisi yang lain, persamaan autoregressive adalah persamaan dimana nilai pada suatu waktu y_t bergantung pada nilai pada lags sebelumnya y_t-p. Persamaan matematis model AR orde ke-p adalah

y_t = μ + φ₁ y_t−1 + φ₂ y_t−2 +· · ·+φ_p y_t−p + u_t Persamaan (6)

dimana u_t merupakan suku galat.

Adalah hal yang umum dalam pemodelan pergerakan harga dan imbal hasil saham dengan menggunakan model AR dan/atau MA. Dan pendekatan yang diberikan oleh Box dan Jenkins pada tahun 1976 hingga saat ini masih sering dijadikan acuan karena dianggap praktis dan pragmatis. Pendekatan Box-Jenkins dapat dijelaskan sebagai berikut.

Pertama kali, tentukan orde p,q yang dibutuhkan pada pemodelan sedemikian sehingga dapat menggambarkan dengan baik informasi yang ada dalam data. Visualisasi seperti korelogram dan perbandingan acf dan pacf dapat membantu dalam hal ini.

Langkah kedua adalah melakukan estimasi parameter model sebagaimana yang telah ditentukan pada langkah pertama tadi. Estimasi dapat menggunakan teknik least square atau maximum likelihood.

Langkah terakhir –dan merupakan tahapan yang paling penting adalah pemeriksaan model; menentukan apakah model yang dihasilkan cukup baik dalam mewakili data. Box dan Jenkins menyarankan dua metode, yaitu overvitting dan diagnosa galat. Overvitting adalah memperbesar struktur pemodelan melebihi apa yang diperlukan untuk memodelkan data. Apabila penambahan struktur ini tetap tidak mampu menghasilkan parameter yang signifikan, maka model awal tersebut sudah cukup baik. Diagnosa galat adalah melakukan pemeriksaan pada karakter yang ada dalam deret galat, apakah sudah memenuhi ciri-ciri white noise atau belum.

Pemodelan Volatilitas

Sejauh ini, pemodelan yang telah dijelaskan menganggap bahwa varian suku galat pada persamaan rerata (5) dan (6) adalah konstan. [2] dan [3] berpendapat bahwa varian ini tidak konstan; tapi berubah-ubah sepanjang waktu sehingga perlu dimodelkan.

Untuk mengetahui apakah suatu deret waktu memiliki varian yang berubah-ubah sepanjang waktu, uji ARCH effect dapat digunakan. Langkah pertama, lakukan estimasi parameter dengan regresi linear seperti biasa. Dari deret residu u_t yang dihasilkan, lakukan regresi untuk model sebagai berikut

u²_t = γ₀ + γ₁u²_{t −1} + γ₂u²_{t −2} +· · ·+γ_qu²_{t –q} + v_t Persamaan (7)

Dengan mendefinisikan TR² sebagai perkalian antara jumlah pengamatan dan koefisien korelasi serta mengacu pada distribusi χ(derajat kebebasan = q), dapat ditentukan apakah H₀: {γ₁ = 0 dan γ₂ = 0 dan . . . dan γ_q = 0} atau H₁ {γ₁ ≠ 0 atau γ₂ ≠ 0 atau ... atau γ_q ≠ 0} yang tidak dapat ditolak.

Salah satu model non-linear yang banyak digunakan di bidang keuangan adalah model ARCH (kepanjangan dari Autoregressive Conditional Heteroscedasticity). Salah satu fitur dari model ini adalah kemampuannya dalam memodelkan leptokurtosis dan volatility clustering. Persamaan matematisnya adalah

y_t = β₁ + β₂x_2t +… + β_nx_nt + u_t dimana u_t ∼ N(0, σ_t²) Persamaan (8.a)

dan σ_t² didefinisikan sebagai

σ_t² = α₀ + α₁u²_{t −1} + α₂u²_{t −2} +· · ·+α_qu²_{t –q} Persamaan (8.b)

Bandingkan persamaan (2) dengan persamaan (8.a) diatas. Pada persamaan (2), suku residu masih berdistribusi normal, namun sekarang merupakan fungsi dari waktu.

Bentuk lain dari pemodelan non-konstan varian adalah Generalized ARCH (GARCH). Model ini beranggapan bahwa varian pada suatu waktu merupakan fungsi dari varian pada waktu-waktu sebelumnya. Model matematisnya adalah sebagai berikut.

y_t = β₁ + β₂x_2t +… + β_nx_nt + u_t dimana u_t ∼ N(0, σ_t²) Persamaan (9.a)

σ²_t = α₀ + α₁u²_{t −1} + βσ²_t−1 Persamaan (9.b)

Persamaan (9) disebut sebagai persamaan GARCH(1,1) karena terdiri dari suku kuadrat galat lag 1 dan suku kuadrat varian lag 1.

Meskipun model yang diberikan pada persamaan (8) dan (9) telah mengakomodasi non-linearitas pada suatu peubah keuangan, kekurangan utamanya adalah model tersebut masih menganggap bahwa gerakan harga saham dan imbal hasil akan memberikan respon yang sama apabila diberikan informasi negatif maupun informasi positif. Di bidang keuangan, fenomena ini disebut sebagai leverage effect. Dijelaskan di [1], apabila harga saham jatuh, maka rasio utang terhadap ekuitas akan bertambah. Bagi investor, hal ini meningkat resiko investasi.

Untuk mengakomodasi ketidaksimetrisan gerakan harga saham dan imbal hasil terhadap informasi negatif dan positif, terdapat dua model utama yang akan digunakan dalam tulisan ini. Yang pertama adalah model GJR, dengan persamaan matematis sebagai berikut.

σ²t = α₀ + α₁u²_t
−1 + βσ²_t−1 + γ_GJR u²_t −1I_t−1 Persamaan (10)

dimana I_t−1 = 1 bila u_t−1 < 0. Bila tidak, I_t−1 = 0. Untuk deret yang memiliki leverage effect, γ_GJR > 0.

Model yang kedua adalah EGARCH, dengan persamaan matematis sebagai berikut.

Persamaan (11)

Pada persamaan ini, ketidaksimetrisan ditunjukkan dengan γ < 0.

Pada pemodelan proses, analisis akan dilakukan untuk imbal hasil di tiga periode, yaitu 2006 – 2007, 2008 – 2009, dan 2010 – 2011. Untuk menguji kemampuan model meramalkan harga saham perusahaan pada sepanjang tahun 2012, akan digunakan model pada periode terdekat, yaitu 2010 – 2011. Selain itu, seluruh analisis menggunakan tingkat kesalahan α 5% kecuali disebutkan lain.

Proses Rerata

Sebelum memulai pemodelan proses rerata, langkah pertama adalah mencari bentuk data saham yang memiliki sifat stasioner. Hasil uji Augmented Dickey-Fuller (ADF) terhadap harga saham perusahaan menunjukkan bahwa data harga saham untuk ketiga periode tidak stasioner dan memerlukan satu kali diferensiasi untuk dapat memulai pemodelan. Diferensiasi sekali ini tidak lain merupakan imbal hasil saham.

Table 1: Uji Stasioner Dickey-Fuller, t-critical -3.9758 pada α 5%

Periode	Data Level		Diferensiasi Sekali
Periode	t-Statistic	Prob	t-Statistic	Prob
2006 – 2007	-3.0635	0.1163	-23.3639	0.0000
2008 – 2009	-1.2182	0.9050	-25.2496	0.0000
2010 – 2011	-2.0553	0.5690	-22.4946	0.0000

Struktur utama untuk proses rerata yang akan digunakan adalah Autoregressive (AR) dan Moving Average (MA). Struktur rerata ARMA ini menunjukkan korelasi antara imbal hasil pada suatu hari dengan imbal hasil dan galat imbal hasil pada hari-hari sebelumnya (lag). Signifikansi informasi pada hari-hari sebelumnya dapat ditunjukkan melalui visualisasi korelogram untuk lag yang bersesuaian.

Berikut dirangkum struktur ARMA untuk ketiga periode, sebagai berikut.

Table 2: Persamaan matematis pemodelan rerata. Nilai-p ditunjukkan dalam kurung tutup.

Periode	Model Rerata	Catatan
2006 – 2007	r = -0.0001 + ε (0.80)	Persamaan (a)
2008 – 2009	r = -0.0017 + ε (0.25)	Persamaan (b)
2010 – 2011	r = -0.0003 + ε (0.68)	Persamaan (c)

Dari tabel diatas, nampak bahwa pada periode 2006 – 2007 dan 2010 – 2011, data level pergerakan saham sudah white-noise. Ini berarti, peubah AR dan MA tidak mampu lagi menjelaskan apa ‘yang ada’ di dalam deret waktu. Dengan kata lain, sudah tidak terdapat informasi di dalam harga saham terdahulu yang dapat digunakan untuk memodelkan dinamika saham pada periode tersebut. Sedangkan untuk periode 2010 – 2011, evaluasi awal pada korelogram menunjukkan struktur ARMA yang signifikan memerlukan informasi imbal hasil pada lag ketiga. Namun pada analisis non-linear, ditemukan bahwa struktur tersebut tidak signifikan, menyisakan konstantan C sebesar 0.0003 pada persamaan rerata dan residu tetap white noise.

Proses Varian

Setelah proses rerata dilakukan dan menghasilkan struktur persamaan dengan residu yang white noise, langkah selanjutnya adalah menentukan apakah kuadrat residu u²_t pada persamaan merupakan fungsi dari waktu. Hal ini penting dilakukan mengingat bahwa asumsi yang mendasari pemodelan rerata adalah bahwa varians dari residu tersebut memiliki varians yang tidak berubah-ubah (konstan) dan residu pada suatu hari tidak memiliki korelasi dengan residu-residu pada hari-hari sebelumnya. Apabila yang terjadi sebaliknya, maka diperlukan persamaan tambahan yang dapat menggambarkan secara lebih baik, dinamika dari varians residu σ²_t.

Evaluasi atas ada tidaknya heteroskedastik pada deret kuadrat residu, dapat dilakukan dengan melihat korelogram kuadrat residu (apendiks C). Selain itu, uji heteroskedastik dengan jenis uji ARCH dapat digunakan untuk tujuan yang sama. Perhatikan tabel 3 dibawah.

Table 3: Uji heteroskedastik jenis ARCH. A adalah pemodelan rerata saja. B adalah pemodelan rerata dan varian

Periode	Prob. F(6,508)		Prob. Chi-Square(6)
Periode	A	B	A	B
2006 – 2007	0.0000	0.9526	0.0000	0.9516
2008 – 2009	0.0000	0.9805	0.0000	0.9801
2010 – 2011	0.0000	0.8145	0.0000	0.8119

Tabel 3 diatas membandingkan hasil uji heteroskedastik kuadrat residu dari persamaan rerata saja (kolom A) dan kuadrat residu dari persamaan rerata dan varian (kolom B). Dari hasil diatas –yang juga sesuai dengan analisis korelogram kuadrat residu (apendiks C), terlihat pada kolom A bahwa kuadrat residu pada proses rerata sangat heteroskedastik –ditandai dengan probabilitas F dan Chi-Square yang dibawah α/2. Pemodelan varian dengan struktur GARCH(1,1) untuk periode 2006 – 2007 dan 2008 – 2009 serta GARCH(2,1) untuk periode 2010 – 2011 dapat menghasilkan model dengan residu yang white noise. Kolom B pada tabel 3 diatas merupakan hasil uji heteroskedastik dengan menggunakan model lengkap sebagaimana disarikan pada tabel 4 dibawah.

Table 4: Persamaan matematis pemodelan simetris GARCH. Nilai-p ditunjukkan pada kurung tutup

	2006 – 2007	2008 – 2009	2010 – 2011
Proses Rerata
C	0.000138 (0.739)	-0.001852 (0.103)	0.001023 (0.1246)
Proses Varian
C	7.21E-6 (0.000)	1.99E-5 (0.010)	3.24E-5 (0.000)
α₁	0.08148 (0.000)	0.10838 (0.000)	-0.02570 (0.242)
α₂	N/A	N/A	0.13072 (0.000)
β	0.84176 (0.000)	0.87807 (0.000)	0.79083 (0.000)
Persistence α₁+ α₂+ β	0.92324	0.98645	0.89585
Waktu-paruh ln0.5/ln(α₁+ α₂+ β)	9 hari	51 hari	6 hari

Dari tabel 4, juga dapat disimpulkan bahwa volatilitas saham perusahaan dalam kurun waktu 2008 – 2009 merupakan yang tertinggi, diikuti dengan volatilitas dalam kurun waktu 2006 – 2007 kemudian 2010 – 2011.

Persamaan pada tabel 4 diatas merupakan persamaan simetris, dimana pergerakan imbal hasil akan sama apabila dipaparkan terhadap informasi negatif maupun informasi positif. Sebagaimana telah dijelaskan pada bagian sebelumnya bahwa kenyataan di lapangan menunjukkan peubah keuangan memiliki sifat yang tidak simetris.

Untuk mengetahui apakah terdapat perbedaan perilaku imbal hasil terhadap informasi negatif dibandingkan dengan informasi positif, berikut akan dilakukan pemodelan GJR atau EGARCH (akan dipilih satu, mana yang memiliki AIC terendah).

Table 5: Pemodelan EGARCH. Nilai-p ditunjukkan dalam kurung tutup

	2006 – 2007	2008 – 2009	2010 – 2011
Proses rerata
C	0.000222 (0.5905)	-0.001599 (0.1597)	0.000443 (0.5373)
Proses varian
ω	-0.707400 (0.0003)	-0.293108 (0.0002)	-0.644634 (0.0000)
α	0.148612 (0.0001)	0.165787 (0.0001)	0.096953 (0.0112)
γ	-0.032900 (0.1756)	-0.065104 (0.0037)	-0.124790 (0.0000)
β	0.936106 (0.0000)	0.975645 (0.0000)	0.929810 (0.0000)
AIC	-6.487485	-4.233948	-5.291227

Dengan membandingkan Akaike’s Information Criteria (AIC) untuk model EGARCH (tabel 5) dan model GJR (tabel 6) pada masing-masing periode, dapat disimpulkan bahwa model EGARCH lebih dapat merepresentasikan dinamika imbal hasil pada periode 2008 – 2009. Sedangkan pada periode 2006 – 2007 dan 2010 – 2011, model GJR lebih dapat merepresentasikan dinamika imbal hasil perusahaan.

Table 6: Pemodelan GJR. Nilai-p ditunjukkan dalam kurung tutup.

	2006 – 2007	2008 – 2009	2010 – 2011
Proses rerata
C	9.38E-5 (0.8203)	-0.002107 (0.0814)	0.000641 (0.3716)
Proses varian
α₀	6.9E-6 (0.0005)	2.39E-5 (0.0077)	1.57E-5 (0.0000)
α₁	0.058756 (0.0180)	0.072529 (0.0172)	-0.016021 (0.1749)
γ	0.037171 (0.3254)	0.061966 (0.0599)	0.118690 (0.0000)
β	0.849349 (0.0000)	0.876799 (0.0000)	0.902811 (0.0000)
AIC	-6.491290	-4.217983	-5.294077

Peramalan

Mungkin hal terpenting dari pemodelan dinamika peubah keuangan adalah bagaimana model matematis tersebut dapat membantu memperkirakan pergerakan peubah keuangan di masa yang akan datang. Kerangka kerja utamanya adalah dengan menggunakan harga saham dan imbal hasil pada rentang waktu sebelumnya, dapat diramalkan bagaimana kemungkinan pergerakan imbal hasil dan harga saham pada rentang waktu di depan.

Dalam hal ini, pemodelan pada periode 2010 – 2011 akan digunakan untuk meramalkan pergerakan imbal hasil dan harga saham sepanjang tahun 2012 (sebenarnya tidak benar-benar diramalkan, karena tahun tersebut juga telah lewat). Model yang akan digunakan adalah model asimetris GJR (Tabel 6 kolom paling kanan) karena lebih mendekati kenyataan yang ada pada tahun 2012 tersebut, dimana unsur leverage sangat signifikan. Hasilnya adalah sebagai berikut,

Gambar 1: Model GJR menggunakan periode 2010 – 2011 digunakan untuk meramalkan pergerakan imbal hasil di 2012. Area yang diarsis adalah rentang ±2SE

Dari pergerakan imbal hasil pada gambar 1, pergerakan harga saham dapat diperoleh dengan cara melakukan integrasi sekali (karena diawal analisis telah dilakukan diferensiasi sekali) terhadap model asimetris GJR tersebut. Secara empiris, EViews dapat mengeluarkan deret imbal hasil dan galat standar ramalan, karena itu proses integrasi dapat dengan mudah dilakukan. Hasilnya adalah sebagai berikut.

Gambar 2: Perbandingan gerakan saham aktual dan ramalan

Dari gambar 2 diatas, terlihat bahwa model GRJ yang menggunakan sampel 2010 – 2011 cukup baik dalam meramalkan gerakan saham di tahun 2012, utamanya pada periode Mei – Juli 2012 dan Desember 2012, kecuali perbedaan yang sangat mencolok disekitar April dan Oktober 2012. Perlu dicatat bahwa ekspektasi rerata imbal hasil sepanjang 2012 adalah berupa konstanta sehingga hasil integrasinya adalah pergerakan harga saham yang merupakan garis lurus dengan kemiringan sebesar +0.000641. Area yang diarsir merupakan batas atas dan bawah sebesar dua kali galat-standar.

Diskusi

Beberapa hal patut untuk didiskusikan disini. Sebagaimana telah dijelaskan pada bagian pendahuluan, bahwa proses pemodelan peubah keuangan, dari proses rerata hingga proses varian merupakan proses yang (bisa jadi) berulang-ulang. Saat struktur arma telah ditentukan dan memberikan residu yang white noise, ada kemungkinan struktur ini tidak dapat digunakan pada pemodelan varians tanpa memunculkan korelasi pada proses rerata dan/atau heteroskedastis pada proses varian. Sebagai contoh pada periode 2010 – 2011, analisis korelogram imbal hasil menunjukkan hubungan pada AR(3) dan MA(3). Namun pada saat pemodelan varian, residu yang white noise (dalam hal ini, sudah homoskedastik) hanya dapat dihasilkan apabila struktur AR dan MA tersebut dihilangkan dan menyisakan konstanta.

Untuk ketiga rentang waktu yang dianalisis, proses reratanya seluruhnya menunjukkan adanya heteroskedastisitas pada residu. Dan dengan menggunakan struktur GARCH, memang tampak bahwa varian residu tidak konstan dan merupakan fungsi dari residu dan varian itu sendiri. Terlihat pada tabel 4 diatas bahwa periode 2008 – 2009 menunjukkan volatilitas imbal hasil yang sangat tinggi –dengan waktu-paruh selama 50 hari, sangat tinggi dibandingkan dengan periode sebelum dan sesudahnya, masing-masing 9 hari dan 5 hari. Pada rentang waktu tersebut, dunia sedang mengalami krisis keuangan global, khususnya di bidang perbankan di Amerika Serikat.

Kenyataannya adalah bahwa setengah dari keuntungan yang diperoleh perusahaan di tahun-tahun sebelumnya berasal dari General Electric Capital, anak perusahaan yang dimiliki 100%, dan melakukan pendanaan hampir keseluruh jenis layanan keuangan, mulai dari penyewaan mobil di Eropa hingga investasi perumahaan komersil di Florida, Amerika Serikat. Perusahaan ini juga memiliki aset hampir 1,800 pesawat komersil dan menyewakannya ke sekitar 225 perusahaan penerbangan.

RUJUKAN

Brooks, C., “Introductory Econometrics for Finance, Second Edition”. 2008. Cambridge University Press.
Luetkepohl, H., “New Introduction to Multiple Time Series Analysis”. 2005. Springer-Verlag
Nicholas, A., Sophia, E., “Stock returns and volatility: Evidence from the Athens Stock Market Index”. 2001. Journal of Economics and Finance

Investment –How to Allocate Money into Equities

2014-02-01T22:29:00.001+07:00

Hi there.

As the investment manager always says, the most crucial step on divvying money into investment market has been how much fraction of money should be put between risky assets. Now I would like to initiate discussion –practical guidance on how to invest money into the equity market.

In this discussion, the primary objective is the highest return on investment in the long term of investment period. Furthermore, it is how to achieve optimum return while maintain minimum risk (and apparently this is the most celebrated Sharpe ratio) using historical return data from the past. This would be a naive structure of analysis and hence discussions are appreciated. I would leave the exact definition of long term for discussion.

Commonly, the first step is a qualitative evaluation to select what equity should be purchased. I would like to skip the evaluation and give the already chosen equities. The figure below shows the relation between index return of Bursa Efek Jakarta (horizontal axis) and equity return of each company (vertical axis), comprises of data from January to September 2013.

Five equities are hence selected for the risky portfolio from a broad of industries, i.e.: Inco (INCO), Unilever (UNVR), Astra International (ASII), Tambang Batu Bara Bukit Asam (PTBA), and Bank Rakyat Indonesia (BBRI). These are all listed in Bursa Efek Jakarta. It is apparent that return of ASII is very positive sensitive to the dynamic of return of the market while its risk is very low –represented by distribution of the return points that are close to the trend line. At the other side, dynamic of return of INCO is slightly negative to the dynamic of return of the market while its risk is very high –distribution of the return points is spread away from trend line. Preliminary suggestion would advise that investor should put the money away from INCO and PTBA as the risks are very high!

Statistic values extracted from the historical data of market and equity return and used for our analysis are as follows.

Variable beta represents sensitivity of equity return and market return. Variable alpha represents how much equity returns while the market return is zero. Two other important statistic variables are residual variance and residual sum of squares. It is noteworthy the “i” denotes individual equity and “p” denotes collection of equity or the portfolio.

Technical assumption and constraint for finding optimal money allocation are as follow.

It is assumed that short-sale is prohibited. Thus, allocation would not be negative,
All money shall be invested into five equities. Thus, sum of the fraction of allocation is 1, and
Minimum allocation is 1% of total money

Statistic values of portfolio are calculated as follow,

Alpha p: sum of product of Alpha i and its allocation
Beta p: sum of product of Beta i and its allocation
Residual value p: sum of product of residual value i and square of its allocation
Return p: alpha p + beta p * (expected return of market)

Adjusted variable is calculated as follow,

Standard deviation p: square-root of (variance of market return * square of Beta p + residual value p)

And objective variable is calculated as follow

Sharpe ratio p: return of portfolio divided by its standard deviation

Using Excel spreadsheet, these four aspects are workable using Solver module.

Now, the procedure for calculations is

Formulate assumptions, constraints, and calculations mentioned previously into spreadsheet,
Set the equity weight wi 0% for INCO, UNVR, BBRI, ASII and leaves PTBA 100% allocation,
In the Solver module, set the target value to maximize Return portfolio, by changing values of equity allocation and subject to the assumptions and constraints, minimum allocation of 1%, sum of allocation is 100%, and portfolio standard deviation is 0.017. Don’t forget to click “Solve” button.
Excel will then provide you with the result of each equity allocation. Write down the resulted expected return of portfolio paired with the standard deviation of portfolio

Repeat the procedure for another value of portfolio standard deviation and the corresponding portfolio return would be such this figure below

The blue line is one of the most well-known curve in the theory of investment; efficient frontier made famous by Markowitz. The line represents a set of maximum expected return that would yield from specific allocation into risky assets for each value of risk (represented by standard deviation) an investor are willing to take.

As with our main objective, the highest Sharpe ratio is indicated by red dot from the efficient frontier line above. With a 0.016 standard deviation, the maximum expected return from portfolio would be 0.3%. It means 0.178 Sharpe ratio. And this also relates to the following allocation of money.

From the table, INCO and PTBA were historically poor performer. Meanwhile, more than two-third of money should be invested into BBRI.

Now is for the practical consideration. Equity selection involves qualitative analysis of how is the company business plan, who is the CEO and its management team, etc. Thus, it should be long-term activity.

Dive into more frequent activity, as the equity price of a company is changing rapidly; the money allocation should also change over time to keep the Sharpe ratio maximum. Therefore, investment manager should frequently extract historical data and initiate the calculation procedure. With a certain amount of subscription fee, Bloomberg provides an add-on to simplify data extraction into spreadsheet. Thus, calculation for money allocation could be scheduled for each week or even each hour.

Note: please find spreadsheet of calculation here. And as always, discussion are welcome.

Contribution of State Owned Company for National Economic Growth

2014-01-14T19:24:00.001+07:00

Hi there.

Now I would like to share an article that was submitted few months ago for my graduate scholarship in Universitas Indonesia. Nevertheless, I am not confident about the calculation (you could download it here) so please give me feedback. Noteworthy is all data and values were extracted from annual reports of the company, which you could download it here.

Since 2005, Garuda Indonesia group began a program of rehabilitation in order to boost its financial condition. The need to change the internal work culture and the challenges of regional economic conditions that had drawn in 2008, it is slowly but surely, the corporate income was improved. It can be seen at the company's revenue contribution to the economic growth of the nation. Note the picture below.

In general, the company's contribution to the nation's economic growth is quite good, between 0.3% to 2.7% except in 2006 and 2009 with a negative contribution due to the decrease in operating revenues. These decreases resulted from the consolidation program executed by management in order to ensure continuity of company production level in the long run.

In 2006, the company decided to stop the negative cash flow by phasing-out dozens of classic aircraft and in a timely manner began to operate new generation aircrafts. In 2009, the competition from low cost carriers are somehow very tight resulted in the company's domestic market share eroded to 27.4% from 29.5% in the previous year even though the company had been trying to reduce passenger fare yield. The result was a decrease in total revenue.

By improving the performance of production level and maintain a strict financial sustainability, it is expected that contribution of Garuda Indonesia is significantly increasing to the nation's economy.

Good is not good when better is expected - Mr. Emirsyah Satar, President Director of Garuda Indonesia.

Aircraft Appraisals

2013-12-23T22:26:00.002+07:00

Hi there.

First of all, this is my first post after two years. I am not gonna post something about flight data recorder as my previous post. Instead, I would like to share with you, another interesting topic in airline industry.

Many aviation-related-companies in this world offer an aircraft valuation services. For a specific condition of aircraft and in a particular time reference, they would provide you with -let me say an opinion from certified person about what the dollar value of your aircraft(s) should be, in worldwide market.

Few of many prominent non-profit community that provide certification for aircraft appraisals are International Society of Transport Aircraft Trading and Institute of Business Appraisers. And as far as i know, company such as Ascend, IBA, and Morten Beyer and Agnew Appraisals have at least one of their employee certified from one of these institutions.

An opinion from certified appraisers is of course a very important reference for insurance broker and aircraft owner. Insurance broker calculates what amount of risk premium that airline shall bear to insure the aircraft. This premium is basically depend on the aircraft value. An aircraft owner use the opinion to determine how much dollar money shall the airline (who operates his/her aircraft) pay to him/her should there is total loss involving the aircraft.

Okay, I am gonna stop the introduction. What I am gonna share with you is a -well, at least based on my experiences, method and formula on how to fairly calculate an aircraft value. In this example, I would like to value an Boeing 737 aircraft in its middle age of economic life (had done airframe and engine heavy maintenance). Let us assume that this aircraft is under MSG-2 block-check maintenance.

Valuation would be divided into two main parts. First, calculation of a so called adjusted value due to maintenance status. A non certified person could do this step. Second, final calculation involving an economic-based estimation of aircraft and engine half-life value which should only be done by certified and experienced person or entity.

Adjusted valuation due to maintenance status of an aircraft would require the following cost information:

Last D-Check date, time since Last D-Check, and its maintenance interval
Last landing gear overhaul date, time since last overhaul, and its maintenance interval
Last engine refurbishment date, time since last refurbishment, and its shop-visit interval estimate
Last LLP replacement date (for each LLP module), time since last replacement, and its replacement interval
Last APU overhaul date, time since last overhaul, and its shop-visit interval estimate

In addition, one would also need the following aircraft and engine maintenance status:

Total cost of D-Check
Landing gear overhaul cost
Engine refurbishment cost (refurbishment to engine full EGT margin)
Engine LLP replacement cost
APU Overhaul cost

Maintenance status and cost information. Numbers are for illustrative only.

Engine LLP status and cost information. Numbers are for illustrative only.

From the first table above, the airframe last D-Check was on March 2005 and it costs $900,000. Aircraft flight hours was used as timer for this D-Check maintenance and its interval is 30,000 flight hours and until the calculation was prepared, the aircraft had been flying for 25,000 flight hours. The landing gear was overhauled on January 2007 and it costs $100,000. The interval is 120 calendar months and had been used for 100 months.

APU was overhauled on February 2009 and costs $50,000. The interval was estimated for 36 months and had been used for 20 months. Two engines was refurbished concurrently on May 2006 and cost $1,000,000 for each engine. The interval was estimated for 25,000 flight cycle and had been used for 24,000. The second table was used to calculate adjustment value for LLP of engines.

If you observe the usage of the parts mentioned above, you may notice that most of the maintenance had been 'used' by more than half of its economic life, except the D-Check which still have 25,000 of 30,000 flight hours remains. This lead us to an assumption that the adjusted value would be relatively low compared to its full maintenance value.

Now, for the formula. The adjusted value for each parts should be a fraction of its remaining maintenance value to its full maintenance value, multiplied by maintenance cost, i.e.:

AV = (0.5 - (Time Since Last Mtx / Mtx Interval)) * Mtx Cost

Formula above was self-explanatory. If the aircraft was fresh from D-Check, then time since last maintenance would zero and its adjusted value would be equal to half of its maintenance cost. Or if the engine EGT margin would go negative in a few days of flight, then its maintenance fraction would be almost equal to minus zero point five (for example: 0.5 - 24,990/25,000) and makes it trash of aluminium!

Using the two tables above and calculating the adjusted value by formula already mentioned, then one would arrive at -$545,800 of adjusted value. This negative dollar value had been guessed earlier, as most of parts had been used for more than half of its economic life. For clarity, I share your my excel calculation here.

Now that the adjusted value obtained, the final step is required. A good definition of half-life value of an aircraft is given by ISTAT,

'Half-life' condition assumes that every component or maintenance service which has a prescribed interval that determines its service life, overhaul interval or interval between maintenance services, is at a condition which is one-half of the total interval.

In general, half-life value of an aircraft was dependent on aircraft age, aircraft type, and aircraft maximum take-off weight. Assuming that one of those certified appraiser state that half-life value of the Boeing 737 aircraft with its age and characteristic is $5,000,000, then our final aircraft value would be $5,000,000 + (-$545,800) or equal to $4,454,200. End.

Dear readers, I would be very much thankful should you have time to give me feedback on my thoughts or would like to drop any comments. See you on another interesting topic on aviation.

FDR Readout Free For Public

2011-06-16T16:55:00.123+07:00

hi there.

It's been about a month since the first post on FDR readout using Matlab. Now, allow me to show and to give you the application. I called it aPENGAMAT. In Bahasa Indonesia, this is Analisis Penerbangan dan Keselamatan. In English, it could be Flight Safety Analysis. Yes I know, the name of the software is quite a big deal while this initial release is no more than a very simple readout process. I have a dream (well, only a dream for now) that this software could also be used as a real flight safety analysis such as AirFASE. In fact, software routines are basically mimicking AirFASE. OK, let me introduce you. And at the end of this article, I'll give you the software ;)

Since this is based on Matlab 2008a Engine, you will need to instal Matlab Compiler Runtime (MCR 7.8). It's funny enough that I could not find download page in Mathworks site!

Edited: It seemed Mathworks would not allow any public reference for downloading it. You may need to purchase a Matlab license to obtain MCR included. Thank you Sumandari for MCR Installer link.

If you click the executable file, the main windows would be like this.
Just like the AirFASE routine goes on, there will be 1+3 simple step to be able to read FDR file. At first, you will need to setup -a so called dataframe layout that is customized for each aircraft. Ask for it to airline or manufacturer of the aircraft. May be you are wondering now, what is the dataframe layout? Well, you can read a general explanation here and here. If you have access and/or capability, just go to ARINC site and purchase ARINC 717 specification document. It will cost you a couple hundreds US dollars ;)

But let me give you one quick example what is the dataframe layout and how to define a flight parameter conversion into database. Let me pick a parameter called HEADING ATTITUDE. An aircraft's heading is the direction that the aircraft's nose is pointing. It is referenced by using either the magnetic compass or heading indicator, two instruments that most aircraft have as standard. Using standard instrumentation, it is in reference to the local magnetic north direction. True heading is in relation to the lines of meridian (north-south lines). The units are degrees from north in a clockwise direction. East is 90, south is 180 and west is 270 degrees. Operational range would then be from 0 degree to 360 degree.

Now, from a dataframe layout document, it is stated that the HEADING PARAMETER is stored in word 28, all subframes (1-4), and taking up 10 of 12 bits, ie: bit 3 (LSB) to bit 12 (MSB). See picture below. This parameter location is marked blue.

Now, the conversion function from raw data into engineering unit would be:

EU(DEGREE) = operational_range/bit_len^2 * RAW(DECIMAL)

From the figure above,

RAW = 1111100100 (BINARY) = 996 (DECIMAL)

Knowing the operational_range is 360 degree and bit_len is 10, the engineering unit would be:

HEADING (DEGREE) = 360/10^2 * 996

HEADING (DEGREE) = 0.08 * 996

HEADING = 350.156 DEGREE

This generic rule also valid for some flight parameters, such as pitch and roll attitude. Read the rest of the aircraft dataframe layout document. Now, we are ready to input it into DFL database. See figure below.

To be able to decode the FDR file, this program will try to communicate with dataframe layout database. And since I only have knowledge on Microsoft Access, we will keep in-touch with this closed-source software. To connect aPENGAMAT with Access database, open your Control Panel -> Administrative Tools -> Data Source (ODBC). Obviously, it is for Windows only.

Click "Add" to create new data source -> choose "Microsoft Access Driver (*.mdb,*.accdb)" -> click "Finish".

In "Data Source Name", type DFL. It must only be DFL though. In Database group click "Select...". Now, select DFL.accdb (this file is given. scroll trough the end of this article to download). Click "OK." Now, our dataframe layout database is already registered to Windows. Well done!

Now, we need to build m-file dataframe layout. From main window, click "Build DFL". Depending on your dataframe layout database, a input dialog will appear.

Choose a related dataframe that you want to use to read the FDR file. Then, we need to build -a so called intermediate file. This is no more than a mat file which is already structured in frame-subframe-word as specified in ARINC 717. Click "Build Intermediate" to proceed.

OK. Now, take a look at "Word: 64". It is the value of word per second that is effective to FDR file. Other value would be 128, 256 and 512 wps. Recent Boeing aircraft such as Lion Air B737-900ER with a tail number PK-LHM has used the latter value.

For "Posisi Baca", selecting EoF will read the FDR file starting from the end of file. Otherwise, selecting BoF will read the FDR file starting from the beginning of the file. For "Panjang Bacaan", set the value of "Jam" for hour and "Menit" for minute. Typical single flight would be an hour -or 60 minutes. Here, we are now ready to create intermediate file. Click "Open Raw".

For example, there are two ARINC 717 compatible file with extension of ddf. Selecting one of the files would get you back to Interm window. Now click "Create Interm". Program will now creating the intermediate file. Just wait untill a save window appears. Set the filename and click "Save". Well done!

Step two. From the interm_file.mat which contains ARINC 717-structured data, the program need to build -a so called result file, which also is a mat file that is already contain parameters value in engineering unit (for example, Pressure Altitude in Feet, IAS in Knot, Autopilot status in ON/OFF, and so on). On the main window, click "Build Result".

We are now asked for intermediate file that we had previously created. Open it. Then, a second window is asking us to select a dataframe layout file in mat extension.

Program will now converting the specified raw data into a more usefull information by using specified dataframe layout.

Now, result file had been created. To show result, click "Plot Result".

Go to Option -> Load and select the recently created result file. And there you go, program will try to show all active parameters. Change pop-up menu to show other parameters.

If you want to print these four subplot together, go to Option -> Print. It will be saved to where aPENGAMAT executable file is located.

Now, as my promise, I will give you the files. Here is for dataframe layout access database and here is for the aPENGAMAT executable file. Should you have discussion, please contact me.

Update Log:
23 August 2011,

Redesigned: result file restructured to adapt event finder logic. You can also use this file to show flight parameter into Baca Window.
Access database restructured.
Added: zoom capabilities.

Added: Find Events button. Note: Still struggling coding for the logic ;)

11 August 2011,

Solved: Cannot collect parameter with more than 1Hz sampling rate
Note: I am currently working to add event exceedance detection feature between my idle work hours. Be patient ;)

19 July 2011,

Plot Result Figures (Baca Window) can now be zoomed simultaneously.

13 July 2011,

Intermediate process now runs twice faster!
Teledyne and Hamilton synchronizatation type now available.
64 and 128 wps file can now be read.

Sync Facebook Contacts with Android

2011-05-15T22:07:00.009+07:00

Hi there.

It's been about a month that i realized my Android phone cannot be synched with Facebook account since this phone doesn't have Motoblur account. I am really sad about it because i believe, there will be a time when i need the email address.

So i googled this site, and then found a very simple method on how to synch your Facebook account into your Android contact without software or installation method. Here they are:

1. you must have both yahoo mail and gmail account (gmail account that you use in your android -Harapan).

2 login to your yahoo mail.

3.go to contact - then choose import email from facebook (it suck all your friend email address into your contact list)

4. export yahoo csv. download/save it.

5.login to your gmail (i found that you need to open you gmail account on standard mode as HTML mode doesn't provide you with uploading contact -Harapan)

6. upload the csv file to gmail contact.

I added this one:

7. If you have already register your gmail account on your Android phone, tap your gmail account and tap Sync Contacts.

This trick goes to: www.google.com/profiles/mrashdi

FDR Readout

2011-05-02T10:42:00.014+07:00

Hi there.

Well, since working in airline industry, my Symbian journey is stagnant. I really like it though. It is because, creating application using Symbian compatible mobile device, the limit is the sky. Hehe, yes i mean it.

Now, been working closely with Flight Data Recorder System, I have some knowledge on communication structure used on aircraft. In essence, it has so many sensors around its body. For example, it has accelerometer sensor, attitude sensor, velocity sensor, ect. Why do we need these all? Well, for a reason, aircraft automatic flight control needs it for their feedback system. By acquiring aircraft attitude, the control system can decide wether the aircraft bahaves correctly or not while maintaining it.

Ok. The values sensed by sensors then acquired by a so called Flight Data Acquisition Unit (FDAU) or something like that. By FDAU, all the values is then encoded with ARINC 717 specification before sending it to FDR (some people may say it as blackbox, eventhough the color of FDR is orange).

So, it's now the important part: the collected flight data is now encoded in FDR. But how can we decode it? How can we able to view all or certain flight history, such as altitude, velocity, or attitude?

Well, I would answer it with... you need to read the ARINC 717 specification. All you need to know before able to read/decode FDR raw data is in it. I, for a couple of weeks also try to be able to read this FDR raw data.

This time, because of its nature and for some personal preferences, I chose Matlab as my development tool. It is basically pretty simple. There is one main Matlab function that is the crux of this script, i.e.:

fread(rawId,bitLen,nBit)

For this purpose, set bitLen to 1 (integer) and nBit to ubit1 (string).

From there, you can choose whatever scheme you would like to obtain, to re-structure, and to convert the value. Oh yes, I forget to tell you that because the FDR raw data is encoded in binary format, able to understand ARINC 717 is not enough. There is also a so called Interface Control Document which describe where -let say the sensed accelerometer is stored by FDAU in a dataframe. I think I'm gonna describe it later on different article. Please read the ARINC 717 specification for definition of this dataframe.

Okay, i think it's enough for now. Gotta do something else. Should you have any related discussion, please email me or comment this article.

Jurnal of Aerospace Engineering

2011-05-01T12:10:00.005+07:00

Hi there.

Well, actually this paper was submitted few months ago. You can check it here or here. Abstract is here:

This paper presents a high‐performance and simple guidance method for a short‐range missile. Missile control systems are designed along the estimated flight envelope of the missile based on coefficient diagram method. The LTI models are obtained by linearizing missile nonlinear model at specific operating points. The governing equations of motion are accurately modeled by taking into account the effect of the mass rate and the center of gravity shift. A novel algebraic approach called Coefficient Diagram Method is used to design normal acceleration and roll angle control systems. Proportional navigation law is considered as the guidance law. One case of engagement scenarios is studied and the missile performance is evaluated. The robustness of the proposed controller is tested againts parameter uncertainties and wind disturbance.

And you can download the portable document here. Should you have discussion, email me.

Taste Swype With Bahasa Indonesia

2011-05-01T10:50:00.002+07:00

hi there.

recently, been busy working. um, this time, I would like to share some -which i believe would be useful information on how to type easily on Android set.

have you ever heard Swype input method? Now, It took 2 months to realize that this kind of input method is very useful and prompt. And it also took to a question whether I could use swipe for my own local language, Bahasa Indonesia.

After searching here and there, oficially, there is a hacking apk or a blog that show us how to include Bahasa Indonesia database for Swype. But, here somehow I don't like/want to proceed this way. As you need to root your android. hehe, ga lah yau, ini barang masih garansi, tar kalo ada apa-apa, gimane?!

And then, I choose to find database of Bahasa Indonesia's words (well, you need to unzip the source files and find dict file) and put it into custom Swype word database. But it then came to how to put this database all in one click?! Sorry, I gave up!

So, how can I still use this Swype capability for Indonesian language. Somehow I read the help section on Swype and it says something like these:

"If you want to add words on your custom Swype database, just select all the words you want to add, and then click Swype logo on the virtual keyboard ".

Now, where could we find a bunch of words of Bahasa Indonesia. Well, it's easy if you know. Just go to www.kompasiana.com or any other site that uses Bahasa Indonesia as their main language.

From those sites, just copy all words from their article. Of course it must be in Bahasa Indonesia. Then, open your Android New Email Message, paste these words, select all of it, the Swype keyboard shows up and then push the Swype button. There you go, we now have our own custom words in Bahasa Indonesia. If you need more, repeat this paragraf to add more words.

Uang lembaran 100 rupiah 1992

2011-04-24T12:10:00.001+07:00

saya ada uang kuno (yah, ga kuno-kuno banget sih) keluaran 1992. uang lembar 100 rupiah sebanyak 22 lembar. yang menarik adalah nomor serinya berurutan, diakhir dengan XXXXXX900.
kalo ada yang minat dan sesuai, saya ingin jual saja. hubungi saya aja. tx.

My Paper: ROBIO 2008 (Guidance and Control Synthesis for a Short-range Homing Surface-to-air Missile)

2010-07-26T20:28:00.002+07:00

Hi, 2 years ago i submitted a paper in IEEE event, ROBIO 2008. Click the following link.
Paper-Robio-2008-Bangkok-final.pdf

Its about classical control, proportional guidance, and navigation of a simple hypothetical missile. May be later, I will discuss it here.

Debugging (a more elegant way)

2010-07-26T20:26:00.001+07:00

OK, this way of debugging is more elegant but can't see it working in a real device. So it's only for S60 2nd emulator. Create an empty file in [EPOCROOT]\winscw\c named ErrRd. Please note, no extension.

for example
C:\Symbian\8.1a\S60_2nd_FP3\Epoc32\winscw\c\ErrRd

and there you go. Run your app in emulator, and if there is an error, the emulator will raise a pop up note including the panic code. See the SDK documentation for the explanation of this error. Happy coding.

How to Debug Symbian C++

2010-07-26T20:24:00.002+07:00

Hi guys/girls.

Um, this time i would like to share; how to debug your Symbian C++ code. Actually and as the most right and elegant way to do debugging, you need to put a breakpoint to your suspicious code line. But, i don't really like the way Carbide C++ 1.2 do it. So this is my way.

Suppose you have these lines of code

code1;

function1(code1);

code2=code1;

If the application crashed, and you think that function1 is wrong, create globalnote and display it before and after the function1.

code1;

CAknGlobalNote* globalNote = CAknGlobalNote::NewLC();
globalNote->>ShowNoteL(EAknGlobalInformationNote,_L("Before"));

function1(code1);

globalNote->ShowNoteL(EAknGlobalInformationNote,_L("After"));
CleanupStack::PopAndDestroy(globalNote);

code2=code1;

So, if Before message is popped-up but there is no "After" message popped-up, you're right, there is something wrong with function1. Happy coding.

PS: make sure you include aknglobalnote.h at the beginning of your implementation file.

#include
.
.
.

Find your SIM Card IMSI number

2010-07-26T20:21:00.002+07:00

Hello.

Few days ago, I am in a need of IMSI of my SIM card. If you'd like to know what is IMSI, see this. So, i build a small Symbian based application, compatible with S60 2nd FP2 (haven't try it on other version).

Just install the following app, and open it. There are only two function, ie; getting IMEI and Phone model and getting IMSI number. Hope it helps.

CobaDevInfo.sis

First step to Symbian C++

2010-07-26T20:04:00.002+07:00

Hello.

This is my first post on this blog. Um... I'd like to share a knowledge, how to prep and install Symbian Tools needed to create a Symbian S60 based application. If you haven't know what is Symbian S60, see this. Well, this is C++ language though.

OK, to get started, download S60 SDK also at the link listed above. Then, you need to install ActivePerl. see this. Install it. If you use Windows 7, then the installation won't be able to add environment path. So ad manually. Go to advanced system settings, click environment variables.... Then, on System Variables group, edit the "Path" Variable, add

;C:\perl\bin\

Make sure to add ';' character before the string.

SDK done, ActivePerl done, it's time to install IDE software. You may use Microsoft Visual C++, Warrior. But, i personally prefer Carbide.C++ IDE. The download site is this.

OK, there you go. For the first time, you may need to create a try-out project. Click File on the IDE, then select create Symbian Project. Just follow the dialog, and you'll be good to go.

Ok then, we need to build the project so that it can be installed on S60 phone. First, you need to set the build target to ARMI phone.

Final step. Go to Project | Build Project. You're done. Install it on your S60 Phone.